Hi Oshyan,
This had to come out some day, the day is now
A bit strange for me to read now that the AA-filter paper didn't work out because of a very high post-render calculation cost.
We spoke about it about a week ago and even though I didn't knew the authors/paper title anymore it didn't ring any bell.
Consequently, I don't think it is unreasonable for me to think and say it hasn't been pursued anymore, otherwise a bell would have ringed. See?
Regardless of that, it's good that it has been investigated and it's a pity it didn't work out. That happens.
I think the majority of us is aware about novel technologies possibly not being suitable for implementation.
Otherwise tons of silly references would have already been posted.
However, I strongly disagree with the connection you make between the attempt to integrate the AA-filter (just for example) which didn't seem to be suitable and the fact that other companies also can't jump onto novel technologies because of similar situations.
These companies sometimes even developed those novel technologies themselves.
Nvidia invented CUDA and layed down the path for GPU-based rendering.
Chaosgroup with Vray, but also other companies jumped on the CUDA-train and developed applications at sometimes tremendous speed and with as much financial and human resources as PS, maybe even less.
Besides that if you just compare their oldest versions, they are speed- and quality-wise extremely superior to TG2.
Render with AA12 in TG2 and your trees still look CG/game like and it takes an era figuratively speaking.
So in that regard the raytracing implementation is very basic, rough and perhaps outdated. It certainly "feels"/looks like it's outdated.
"Our" biggest competitor Vue already has the volumetric surfaces, for instance. For years.
Besides that some parts of the renderer, like AA, already have been ported to GPU. It's extremely unlikely they will keep it at AA only.
The majority is also aware about resources, but PS cannot keep up using that as the main reason for lack of development and the main reason why others can do that. (We're all here, as customers, aren't we?)
It certainly has to do with ambition and priorities. And with priorities I do not mean the choices which are made during development-time. I think you know exactly what I mean with this
Response to an early suggestion I made well over a year ago was literally "meh...too much math" which makes it evident for me this is the issue. Whether the response was meant sarcastically or not, in both ways it isn't showing enthusiasm and ambition, nor upright interest.
The reason this pops up now is because... how long is TG2 around? 5 years?
How much has changed, significantly? For 5 years people are doing the same thing, dealing with same workflow (of solving issues, design problems as consequence of workflow/issues).
Some of us (pioneers) here who with Matt, you and many others made this community a damn nice place, actually left TG2 for this.
Interest drops as nothing significant changes which makes TG2 much more pleasant to work with in many different ways.
Something with kids and toys, you see.
"You know you have that too, that once in a while for a couple of months TG2 just doesn't work for you anymore. Always the same issues with lack of freedom in design, the extreme effort for really pleasant images, slow, ugly results despite the tons of epxerience with it" to loosely quote from an e-mail I recieved last week as a reply to my question why I don't see him here anymore.
I can admit I have the same experience/feelings. If TG2 would progress more, faster it certainly would allow me to do more, better, beautiful and thus it would be interesting a lot lot longer.
And yes, we all users have our ideas on how and what
Now you're essentially saying "If you asked 100 users you'd probably get at least 20 different answers as to what should be a priority" and "thus it's more useful we think of our strategy ourselves(, because the feedback we get is not conclusive or helpful (enough))".
And that's just plain wrong if you would ask me.
In the end I think it really has to do with ambition and priorities and not resources alone.
To get back to Vue, the changes they made, how crap some might be technically, they DO it and GO for it. Especially the latter.
Cheers,
Martin
